AI Slop Examples: A Walk Down One Home Page

The clearest AI slop examples are small and sit in order down the page: a leftover tab title, logos nobody can click, a legal page with one line, and buttons and forms that do not do what they promise. This guide walks down one made-up home page, top to bottom, and ties each thing a visitor notices to the area of a product it hurts.

Slop is hard to see in your own work. You know what you meant, so you read your meaning into the page. A visitor only gets what is on the screen, in the order they meet it.

Why walk one page instead of listing patterns?

Because a list lets you nod along and move on. A walk shows how the tells stack up in the order a visitor meets them. We tag each stop with one of the six areas our audit scores, so you can see which part of your product it touches. None of it proves AI made a page. Hand-built sites can show every stop on this walk.

The definition of the word, and the usual first-glance tells like a generic headline, a stock-person hero, or a missing price, live in what AI slop means, and the timed routine for your own site is the 20-minute self-check. Here we stay with the quieter tells that those two pass over.

What does the top of the page say?

Hypothetical example: a made-up home page for an app that turns a podcast episode into show notes. We read it as a first-time visitor would.

1. The browser tab reads "Home." Area: Clarity and Promise. The tab is the first words anyone sees, and it names nothing. It is also what shows up in a bookmark, a history list, and a row of open tabs, so a visitor who comes back later cannot tell your page from the others. A placeholder title is easy to miss, because the page looks finished without anyone touching it. Put what the product does in the tab, in a few plain words.

What do the middle and the footer say?

1. A row of "As seen in" logos, none of them clickable. Area: Trust and Legitimacy. In our sample report, the Slop tells field for this area notes that no fake testimonials, placeholder logos, or generic stock images are used. A row of logos with nothing behind it is the opposite case. A visitor looks for proof before paying, so proof a visitor cannot check can cost you more than no proof. Use real names with permission, or leave the row out.
2. The Terms link opens a page with one line on it. Area: Trust and Legitimacy. Our homepage names a real way to contact you and actual legal pages as signals a customer checks before trusting you with money. A footer link that goes somewhere thin looks worse than no link, because it shows the page was added to tick a box. Have a lawyer help with the legal pages, and add an email you read. This guide is not legal advice.

If the headline, the hero picture, the price, and the footer year also look like they came from a template, our guide to what AI slop means covers those, and our design tips for a vibe-coded app cover the fixes.

What happens after the first click?

1. The "Try it free" button does nothing, or sends you back to the page you were on. Area: Flow and Usability. It reads as slop because the screen looks finished while the one part that matters is missing. A visitor will press it twice, then leave. Wire it up, or take it off the page.
2. The sign-up form accepts an email and shows nothing. Area: Flow and Usability. Did it work? The visitor cannot tell, so they try again or leave. Silence also leaves a visitor wondering whether you got their message at all. Add a plain success message that says what happens next, then test with a real address.
3. A link in the menu lands on an error page. Area: Flow and Usability. A visitor takes that as a sign nobody is watching the site. It often happens after the AI renames or removes a page and leaves the old link behind. Click each link once after every big change.

What hides from a screenshot?

Stops 4 to 6 never show up in a still image. Behavior only appears when something is pressed, so a page can look careful and still break on the first click. That is why our audit does more than look: it reads desktop and mobile screenshots, clicks through the public pages, and may submit forms once with clearly labeled test details.

You can do a small version yourself. It is easy to skip on your own app, because you know where every button is meant to go, so click through once as if you had never seen the page. Then open it on a real phone, because menus, buttons, and forms can shift in ways a desktop view never shows.

The audit reads public pages only, so anything behind a login is out of view, and it is not a security review or a full accessibility audit. For what the AI generator fingerprint check can and cannot tell you, see AI Slop Detector: What It Can and Can't See on Your Site.

Which stop do you fix first?

Whichever sits between a visitor and your main button. Stops 4 to 6 come first, because they decide whether pressing the button works. The tab title in stop 1 costs one minute, so do it at the same time. Proof and legal pages follow, in whatever order your own walk makes obvious. Our report ranks its fixes by impact, written in plain language, so you are not guessing at the order.

When do a few tells add up to a pattern?

When they stack. Nearly every site has a line or an icon that is a little generic, and a single tell is just a to-do item. The picture changes when tells appear down the page, because then the whole thing starts to feel made for everyone.

Our report gives each area a "Slop tells" field. When it finds none, it says so. The sample report does that for the areas where nothing stood out, and its verdict label reads "Real product." The field cuts both ways: it tells you when a site is fine.

Walk your own page the same way, starting at the tab and moving down. Fix a few tells, look again, and the page starts to read as made on purpose.